In April 2000, at a Vermont health centre, Diana Levine was injected with the anti-nausea drug Phenergan in what was supposed to be a vein, but was in fact an artery. The drug caused arterial spasms, which led to gangrene; in the end, Levine's hand and forearm were amputated. Arterial spasming was a known danger with this drug, which was labelled with a warning approved by the US Food and Drug Administration (FDA). Nonetheless, Levine sued Wyeth, Phenergan's maker, in Vermont state court, alleging that the labelling was inadequate. The jury agreed, and the court awarded Levine $6.8 million in damages. Wyeth has now appealed the verdict all the way to the US Supreme Court, which recently agreed to hear the case in October.

The question before the court is whether the FDA's seal of approval ona drug's label should preempt a plaintiff's right to sue the maker on the basis of labelling inadequacies. Such 'failure to warn' claims underpin the vast majority of drug-injury lawsuits. The drug companies maintain that the assessment of FDA experts, who carefully weigh the risks and benefits of a drug before approving its label, should trump the findings of inexpert juries confronted with grievously injured plaintiffs.

The companies have a point. Frivolous or misguided litigation can doom drugs that might have benefited millions. The cost of bringing a new drug to market averages as much as $1.7 billion by one estimate, not least because the companies have to meet a multiplicity of FDA requirements along the way. Fairness suggests that clearing all those hurdles should earn them at least some degree of legal protection — as long as they have shown due diligence in testing their drugs, and have openly declared all the relevant data, both pro and con. Nonetheless, 'some degree' is different from blanket immunity, which the court should avoid for at least two reasons.

First, the industry's argument wrongly implies that science can identify all the risks in advance with absolute certainty. FDA drug approvals are based on clinical trials that are necessarily limited in size and duration, so harmful side effects sometimes won't emerge until a drug is in broader use. Witness the painkiller Vioxx, which was taken by tens of millions of people before it was found to increase their risk of heart attacks and strokes. Removing the right to sue when new side effects emerge would rob most of those affected in the future of redress for their injuries.

Second, the industry's argument presumes that the FDA is poised to update warning labels at a moment's notice. As former FDA commissioner David Kessler said recently, this notion is “unrealistic”. With 11,000 drugs on the US market, and nearly 100 more approved each year, the agency is overwhelmed trying to keep up with new side effects — making lawsuits in state courts an important complement to its regulatory efforts. And, as Kessler also noted, the threat of potential litigation helps to ensure that drug firms are prompt in reporting new adverse effects to the agency and working to update labels.

Supreme Court decisions are notoriously difficult to predict. But it is worth noting that in a recent decision that affects the most advanced, expensive and — in some cases — risky FDA-approved medical devices, the court granted manufacturers exactly the kind of immunity the drug industry is seeking in Wyeth v. Levine. To extend such broad-ranging protection to drugs would be both ill-advised and unjust. Whatever the justices decide, they must preserve the right of future Diana Levines to have their day in court.